Surface defect inspection is a key step to ensure the quality of the hot rolled steel surface. However, current advanced detection (DET) methods have high precision but low detection speed, which hinders the application of the detector in actual production. In this work, a real-time detection network (RDN) focusing on both speed and accuracy is proposed to solve the problem of steel surface defect detection. RDN takes ResNet-dcn, a modular encoding, and decoding network with light weights, as the basic convolutional architecture whose backbone is pretrained on ImageNet. To improve the detection accuracy, a skip layer connection module (SCM) and a pyramid feature fusion module (PFM) are involved into RDN. On the standard dataset NEU-DET, the proposed method can achieve the state-of-the-art recognition speed of 64 frames per second (FPS) and the mean average precision of 80.0% on a single GPU, which fully meets the requirements of the detection accuracy and speed in the actual production line.

The computational methods of protein-protein interaction sites prediction can effectively avoid the shortcomings of high cost and time in traditional experimental approaches. However, the serious class imbalance between interface and non-interface residues on the protein sequences limits the prediction performance of these methods. This work therefore proposed a new strategy, NearMiss-based under-sampling for unbalancing datasets and Random Forest classification (NM-RF), to predict protein interaction sites. Herein, the residues on protein sequences were represented by the PSSM-derived features, hydropathy index (HI) and relative solvent accessibility (RSA). In order to resolve the class imbalance problem, an under-sampling method based on NearMiss algorithm is adopted to remove some non-interface residues, and then the random forest algorithm is used to perform binary classification on the balanced feature datasets. Experiments show that the accuracy of NM-RF model reaches 87.6% and 84.3% on Dtestset72 and PDBtestset164 respectively, which demonstrate the effectiveness of the proposed NM-RF method in differentiating the interface or non-interface residues.

Predicting traffic speed accurately is a very challenging task of the intelligent traffic system (ITS), due to the complex and dynamic spatial-temporal dependencies from both temporal and spatial aspects. There not only exits short-term local neighboring fluctuation and long-term global trend in temporal aspect, but also local and global correlations in spatial aspect. Most existing work focus on the local spatial-temporal dependencies, ignoring the global dynamic spatial-temporal corrections, which is comparably critical for traffic speed prediction. To address this problem, we propose a novel Dynamic Global-Local Spatial-Temporal Network(DGLSTNet) for traffic speed prediction, which consists of multiple spatial-temporal module considering the local and global information simultaneously from both temporal and spatial perspective. Each temporal module applies stacked dilated convolution block to exploit multi-scale local temporal information. Moreover, we empoly a global temporal attention block to capture global dependencies of temporal domain in an attention mechanism. In each spatial module, we not only learn the local but also focus on dynamic global spatial information learned by dymamic graph learning block. Combining the feature results from local and global perspective, the capability and expressiveness of traffic predicting model is improved. Experiment results on two real-world traffic datasets have demonstrated that our proposed model can effectively capture the comprehensive spatial-temporal dependencies and can achieve state-of-the-art prediction performance compared with the existing works.

K-means clustering algorithm is one of the most popular technique for clustering in machine learning, however, in the existing k-means clustering algorithm, the ability of the different features and the importance of the different data objects are treated equally; the discriminative ability of the different features and the importance of the different data objects cannot be differentiated effectively. In the light of this limitation, this paper put forward an enhanced regularized k-means type clustering algorithm with adaptive weights in which we introduced an adaptive feature weights matrix and an adaptive data weights vector into the objective function of the k-means clustering algorithm and we developed a new objective function with l2-norm regularization to the weights of data objects and features, then we obtained the corresponding scientific updating iterative rules of the weights of the different features, the weights of the different data objects and the cluster centers theoretically. In order to evaluate the performance of the new algorithm put forward, extensive experiments were conducted. Experimental results have indicated that our proposed algorithm can improve the clustering performance significantly and are more effective with respects to three metrics: the successful clustering rate (SCR), normal mutual information (NMI) and RandIndex.

Fuzzy c-means algorithm (Fcm) frequently applid in machine learning has been proven an effective clustering approach. However, the traditional Fcm cannot distinguish the importance of the different data objects and the discriminative ability of the different features in the clustering process. In this paper, we propose a new kind of Fcm clustering framework: DwfwFcm.Considering the different data weights and feature weights, an adaptive data weights vector and an adaptive feature weights matrix are introduced into the conventional Fcm and a new objective function is constructed. By the proposed objective function, the corresponding scientific updating iterative rules of the membership matrix, the weights of the different feature, the weights of the different data object and the cluster centers can be derived theoretically.Experimental results have demonstrated that the algorithm proposed in this paper can deliver consistently promising results and improve the clustering performance greatly.